1. Field of the Invention
The present invention relates to photolithographic projection and in particular to a method for testing a polarization state, a method for manufacturing a semiconductor device, and a test substrate for testing the polarization state.
2. Description of the Related Art
When a pattern having a size similar to a wavelength of a polarized light or below the wavelength is formed by lithography, a polarization state of the polarized light affects image formation of the pattern. The polarization state includes an “S polarization” perpendicular to a plane of incidence and a “P polarization” parallel to the plane of incidence. When a line and space pattern is formed, the “S polarization” tends to increase contrast and dose margin more than the “P polarization”. Therefore, in Japanese Patent Laid-Open Publication No. 2004-207709, a polarized light is used to project a mask pattern onto a wafer. A precise pattern is formed on the wafer by the S polarized light incident on the wafer.
The polarized light makes it possible to increase the dose margin when the precise pattern is formed. However, the size error of the precise pattern, caused by the error of the degree of polarization of the polarized light, has recently become a problem. The degree of polarization is defined by dividing a difference between optical intensities of the S and P polarized lights by a total optical intensity of the polarized light. The degree of polarization ranges from minus one to plus one. The error of the degree of polarization affects the image formation of a precise pattern. As the pattern size is enlarged, the affect of the error of the degree of polarization is reduced. Therefore, the error of the degree of polarization is regarded as one factor of an optical proximity effect (OPE). The OPE causes the size error of the pattern projected by a reduction projection system, depending on the size of the pattern. However, there are a plurality of OPE factors in addition to the error of the degree of polarization. Accordingly, it is difficult to identify a single factor of the size error caused by the OPE. Therefore, it is desired to develop a technique to determine the quantity of the polarization state of the polarized light, independent of other factors, when the OPE is observed.